Difficulties in Learning Basic Concepts in Probability and Statistics: Implications for Research

Author(s): Joan Garfield and Andrew Ahlgren

Source: Journal for Research in Mathematics Education, Vol. 19, No. 1 (Jan., 1988), pp. 44-63
Published by: National Council of Teachers of Mathematics
Stable URL: http://www.jstor.org/stable/749110 .
Accessed: 09/12/2013 15:11
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

National Council of Teachers of Mathematics is collaborating with JSTOR to digitize, preserve and extend
access to Journal for Research in Mathematics Education.

http://www.jstor.org

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

Journal for Research in Mathematics Education

1988, Vol. 19, No. 1, 44-63

IN LEARNINGBASICCONCEPTS
DIFFICULTIES
IN PROBABILITYAND STATISTICS:
IMPLICATIONSFOR RESEARCH
JOAN GARFIELD, University of Minnesota
ANDREW AHLGREN, University of Minnesota
There is a growing movement to introduce elements of statistics and probability into
the secondary and even the elementary school curriculum, as part of basic literacy in
mathematics. Although many articles in the education literature recommend how to
teach statistics better, there is little published research on how students actually learn
statistics concepts. The experience of psychologists, educators, and statisticians alike is
that a large proportion of students, even in college, do not understand many of the
basic statistical concepts they have studied. Inadequacies in prerequisite mathematics
skills and abstract reasoning are part of the problem. In addition, research in cognitive
science demonstrates the prevalence of some "intuitive" ways of thinking that interfere
with the learning of correct statistical reasoning. The literature reviewed in this paper
indicates a need for collaborative, cross-disciplinary research on how students come to
think correctly about probability and statistics.

Elements of statistical reasoning have become requisite for a wide range

of fields of study. A survey of offerings at the University of Minnesota, for
example, turned up over 160 statistics courses-of which 40 were introductory-taught in 13 departments. This ubiquity is reflected in the popular
press, where ordinary readers almost daily find reports of medical, economic, or psychological studies-reports that can be understood and evaluated only with some understanding of statistics.
At present very little statistics is typically introduced to students before
they enter college. There is, however, a vigorously growing movement to
introduce elements of statistics and probability into the secondary school
curriculum, and even the elementary school curriculum, as part of the basic
literacy in mathematics that all citizens in today's world should have
(Scheaffer, 1984; Swift, 1982). Statistics and probability were major themes
in recent publications of the National Council of Teachers of Mathematics
(NCTM) (Shufelt, 1983; Shulte, 1981). The American Statistical Association (ASA) and the NCTM, through their Joint Committee on the Curriculum in Statistics and Probability, also have emphasized the desirability of
such a curriculum in several recent publications, including a newsletter, The
Statistics TeacherNetwork. The ASA-NCTM Joint Committee has published
a document with recommended guidelines for teaching statistics within the
K-12 mathematics curriculum (ASA-NCTM Joint Committee, 1985), which
includes rudimentary statistics activities as early as Grades 1 to 3.
The enthusiasm for statistics in the curriculum among specialists in mathematics education is endorsed by the mathematics community in general. In
recent reports concerning the curriculum, the Conference Board of the

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

45
Mathematical Sciences (1982, 1983) described elementary data analysis and
statistics as "more important" than current advanced mathematics topics
and recommended that these topics be included as early as the middle school
curriculum.
Some elements of probability and data description have been included in
major curriculum projects in the past (e.g., American Association for the
Advancement of Science, 1974; Educational Services, 1968). Yet only a few
materials have been written specifically to introduce probability and statistics at the precollege level. In the United States, a set of booklets (Mosteller,
1973) for high school mathematics students was prepared under the supervision of the ASA-NCTM Joint Committee. These numerically focused
booklets are considered "challenging if not impossible" for average high
school students and have not been used extensively (Boruch, Nowakowski,
Zawojewski, & Goldberg, 1984).
In the United Kingdom, the Schools Council Project on Statistical Education has published materials for secondary students covering topics that
illustrate how statistics are used in meaningful contexts in different subject
areas. The emphasis of these materials is on developing concepts rather than
carrying out calculations (P. Holmes, personal communication, 10 April
1985). Several current curriculum development projects, including the University of Chicago School Mathematics Project (1985) and the ASA-NCTM
Quantitative Literacy Project, have produced and tested sets of materials
for students in Grades 7 through 12 (Gnanadesikan, Scheaffer, & Swift,
1986; Landwehr & Watkins, 1986; Landwehr, Watkins, & Swift, 1987;
Newman, Obremski, & Scheaffer, 1986). Yet, despite the enthusiastic development of new instructional materials, little seems to be known about
how to teach probability and statistics effectively. For example, in the introduction to the working-group reports on probability and statistics from
the Fifth International Congress on Mathematics Education (Bell, Low, &
Kilpatrick, 1985), it is noted that many problems still remain in the teaching
of probability and statistics.
In this paper we review the literature related to difficulties that students
at the precollege and college level have in understanding probability and
statistics and discuss the research that is beginning to illuminate the exact
nature of the difficulties, their causes, and what can be done about them.
Our intent is not so much to catalog exhaustively the accumulated findings
as to identify and characterize the nature of the wide range of relevant
endeavors. We conclude with a proposal for a new approach to researchone that is cross-disciplinary and involves experts from several disciplines.
TEACHINGPROBABILITY
AND STATISTICS
College Level
Stochastics as a scientific discipline is usually first taught at the college

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

46

Learning Statistics

level. (We will use the term stochastics to refer to the study of probability
and statistics, as is common in Europe.) The introductory course is usually
divided into three areas: descriptive statistics, probability theory, and inferential statistics (Borovcnik, 1985). The topics typically included in each area
are listed in Table 1. Over the past 20 years, much of the literature on
teaching stochastics has been at the college level. This literature has been
filled with comments by instructors about students not attaining an adequate understanding of basic statistical concepts and not being able to solve
applied statistical problems (Duchastel, 1974; Jolliffe, 1976; Kalton, 1973;
Urquhart, 1971). The experience of most college faculty members in education and the social sciences is that a large proportion of university students
in introductory statistics courses do not understand many of the concepts
they are studying.
Table 1
Typical Content of an Introductory Statistics Course (College Level)
Descriptive statistics
Measures of central tendency (mean, median, mode)
Measures of variability (range, variance, standard deviation)
Measures of position (percentiles, z scores)
Frequency distributions and graphs
Probability theory
Rules (addition, multiplication)
Independent and mutually exclusive events
Random variables
Probability distributions
The binomial distribution
The normal distribution
Samplin
Centrallimit theorem
Inferential statistics
Estimating parameters (mean, variance, proportion, correlation coefficient)
Testing hypotheses

Studies in the research literature confirm this impression. Students often

tend to respond to problems involving mathematics in general by falling
into a "number crunching" mode, plugging quantities into a computational
formula or procedure without forming an internal representation of the
problem (Noddings, Gilbert-MacMillan, & Lutz, 1980). They may be able
to memorize the formulas and the steps to follow in familiar, well-defined
problems but only seldom appear to get much sense of what the rationale
is or how concepts can be applied in new situations (Chervany, Collier,
Fienberg, & Johnson, 1977; Garfield, 1981; Kempthorne, 1980). Within
the conceptual underpinnings, the details they have learned or memorized,
for whatever use they might be, therefore quickly fade.
Recently attention has focused on the processes involved in solving statistical problems (Chervany et al., 1977; Stroup, 1984) and the need for
basing statistics courses on problem solving (Garfield, 1981; Kempthorne,

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

Joan Garfield and Andrew Ahlgren

47

1980). (Courses based on problem solving introduce problems from the

beginning and use real data sets, with statistical concepts being introduced
as the need for them arises.) There does not appear to be substantial evidence
yet for improved practice.
Precollege Level
Most of the recent literature about precollege stochastics instruction,
which is more diverse than the college literature, falls into four categories:
statements concerning the need for instruction at the precollege level
(Pereira-Mendoza & Swift, 1981), suggestions for how to teach at the precollege level (Duncan & Litwiller, 1981; Ernest, 1984; Swift, 1983); descriptions of the curricular role stochastics can play (Richbart, 1981), and
descriptions of the difficulties secondary school students have understanding
the concepts (Carpenter, Corbitt, & Kepner, 1981; Hope & Kelly, 1983;
Shaughnessy, 1981). Only the last category is of interest here.
Research on students' understanding of probability is more extensive than
research on statistics and has developed as an area separately from those
studies cited above, which focus on stochastics in general. There have been
two distinct lines of research on probabilistic understanding: One has focused on school children (e.g., Fischbein, 1975; Green, 1983; Piaget &
Inhelder, 1978), and the other has focused on college students and adults
(e.g., Konold, 1983; Tversky & Kahneman, 1982).
At any level, students appear to have difficulties developing correct intuition about fundamental ideas of probability for at least three reasons. First,
many students have an underlying difficulty with rational number concepts
and proportional reasoning, which are used in calculating, reporting, and
interpreting probabilities (Behr, Lesh, Post, & Silver, 1983). Results reported by the National Assessment of Educational Progress (NAEP) from
the second and third mathematics assessments indicated that students were
generally weak in rational number concepts and had difficulties with basic
concepts involving fractions, decimals, and percents (Carpenter, Corbitt, &
Kepner, 1981; Carpenter, Lindquist, Matthews, & Silver, 1983). Low percentages of students had correct responses to exercises involving complex
concepts and skills requiring understanding of underlying mathematics principles. Difficulties in translating verbal problem statements plague stochastics as they do the rest of school mathematics (Hansen, McCann, & Myers,
1985). Second, probability ideas often appear to conflict with students'
experiences and how they view the world (Kapadia, 1985). We discuss this
conflict below in the section on misconceptions in statistical reasoning.
Third, many students have already developed a distaste for probability
through having been exposed to its study in a highly abstract and formal
way. For this reason, Freudenthal (1973) cautioned against teaching any
technique of "mathematical statistics" even to college freshmen.

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

48

Learning Statistics

FOR OVERCOMINGDIFFICULTIES
RECOMMENDATIONS
Several recommendations from teachers for overcoming difficulties in
learning stochastics can be generalized from the literature cited above.
Teachers should1. introduce topics through activities and simulations, not abstractions;
2. try to arouse in students the feeling that mathematics relates usefully
to reality and is not just symbols, rules, and conventions;
3. use visual illustration and emphasize exploratory data methods;
4. teach descriptive statistics alone without relating it to probability;
5. point out to students common misuses of statistics (say, in news stories
and advertisements);
6. use strategies to improve students' rational number concepts before
approaching proportional reasoning;
7. recognize and confront common errors in students' probabilistic
thinking;
8. create situations requiring probabilistic reasoning that correspond to
the students' views of the world.
The articles cited that recommend instructional methods contain different
mixes of logical argument, teacher intuition, classroom anecdotes, and exhortation on the importance of stochastics topics in schooling. An excellent
collection of this genre is the 1981 NCTM yearbook (Shulte, 1981). These
recommendations are based almost entirely on experience of what has not
worked and speculation about what might work. However, little empirical
research has focused on the effectiveness of the different instructional
methods or teaching approaches in developing statistical and probabilistic
reasoning.
Recently, some research on problem solving has shown that students
receiving deliberate instruction in how to solve problems do become better
problem solvers and are better able to "think mathematically" (Schoenfeld,
1985). What may be needed is similar research on statistical instruction and
students' ability to "think statistically." In an attempt to help students think
statistically there has been some interesting experimentation with the role
of computers in learning stochastics.
USEOF COMPUTERSIN TEACHINGSTOCHASTICS
The increasing prevalence of computers in schools has already had some
influence on the teaching of stochastics and is producing its own literature.
Computers have been used in several ways to aid in the teaching of introductory courses in college. Students may access large mainframe computers
and use statistical packages, such as SPSS (Norusis, 1986) or MINITAB
(Ryan, Joiner, & Ryan, 1976), to do the number-crunching operations for
them (Bialaszewski, 1981) or have problem sets assigned by the computer

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

Joan Garfieldand AndrewAhlgren

49

and then get assistance from the computer in working the problems (Edgar,
1973). Attempts have also been made to use large-system programs to have
students run simulations (Stockburger, 1982). Some mainframe applications
have made use of special graphics terminals to give the students graphic
displays in addition to numerical output (Nygard, 1983). Mainframe computer application is very limited, however, because of the high cost of terminals and a lack of access to special ports.
Microcomputers are more promising than mainframe computers in teaching statistics because they are less expensive and have fairly good graphic
capability. Some recent classroom applications have made use of graphics
to display histograms (Vandermeulen & DeWreede, 1983) or to compute
and display distributions of means of samples drawn from a specific population (Thomas, 1984). Graphics-oriented software is now commercially
available to use in supplementing instruction in introductory courses at the
college level (Doane, 1985; Elzey, 1985). The Quantitative Literacy Project
has also made microcomputer activities an important component of its
materials.
The literature contains suggestions as to how microcomputers in schools
provide opportunities to incorporate stochastics into secondary mathematics courses (Collis, 1983; Fey, 1984). There is, nonetheless, little research
on the efficacy of computers in guiding the design of optimal instruction.
An exceptional research effort to use microcomputer graphics to demonstrate statistical processes to high school students is the design and use of
an Apple II program to illustrate the central limit theorem (Johnson, 1985).
The graphic display of Johnson's program shows samples being taken one
element at a time from a discrete population (one of the several selectable),
indicates the position of the mean for each sample after it is complete, and
moves that mean down to an accumulating distribution of the sample means
at the bottom of the screen. The rationale was that the entire sampling and
distributing story is presented concretely and continuously, with no steps
that the student has to imagine or fill in. Johnson gave microcomputer
presentations to two stratified random samples of students, one group receiving the dynamic presentation and the other receiving a very similar but
static microcomputer presentation. Delayed retention tests after 2 and 6
weeks showed the dynamic presentation group was significantly better at
identifying likely samples and likely distributions of sample means. Johnson
also noted an interesting side effect: Students watching the dynamic presentation showed particular interest in whether successive sample means
would fill gaps or balance asymmetries that remained in the accumulating
distribution, sometimes even cheering when this happened. There was some
evidence that these students may have (by selective memory) come to believe
that such "missing" means are more likely to occur. Subtleties like this effect
demonstrate the need for integrating research and instructional development.

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

Learning Statistics

50

RESEARCHIN COGNITIVEDEVELOPMENT
In general terms, we know what some of the important sources of the
difficulty in learning stochastics are. One already mentioned is students'
level of specific mathematics skills. There is also students' general mental
maturity. A variety of studies of the cognitive development of students in
senior high school indicate that perhaps as many as half cannot think on a
formal operational level (e.g., Herron, 1978; Smith, 1978). For example,
they do not completely grasp proportionality, hypothetical argument, or the
concept of controlling variables (Cantu & Herron, 1978; Green, 1983;
Milakofsky & Patterson, 1979; Stonewater & Stonewater, 1984). Green
(1983) also reports that pupils' verbal ability is often inadequate for accurately describing probabilistic situations.
For many students, a considerable improvement of skills in dealing with
abstractions may be necessary before they are ready for much of the probabilistic reasoning and hypothesis testing that underlie basic statistical inference. For some students, teachers may have to be content to forgo abstraction and to convey what statistical ideas they can in simpler, concrete
terms. It is noteworthy that introductory units in the ASA-NCTM Quantitative Literacy Project have avoided virtually all computations, using instead
plotting and counting methods and using medians rather than means, even
for fitting prediction lines in bivariate plots.
Beyond the underlying skills problems, however, there is an even more
serious source of difficulty: the students' intuitive convictions about statistical phenomena. The second NAEP mathematics assessment produced evidence that students' intuitive notions of probability seemed to get stronger
with age but were not necessarily more correct (Carpenter et al., 1981).
Fischbein (1975) also found decrements in probabilistic performance with
increasing age, which he attributed to school experience and to scientific
reductionism. Students' intuitive ideas, presumably formed through their
experience, may be reasonable in many of the contexts in which students
use them but can be distressingly inconsistent with the statistics concepts
that we would like to teach them. Green (1983) found that performance in
recognizing randomness declines with pupils' age. He hypothesized that two
opposing tendencies are involved: maturation on the one hand and a dominance of scientific deductivism on the other, which "stifles the appreciation
of randomness by seeking to codify and explain everything" (p. 774). As
has been found in recent research on students' intuitive ideas about natural
science, students' ideas cannot be ignored or dismissed. Indeed, it is difficult
to drive such ideas out even with concentrated efforts to do so.
IN SCIENCE
MISCONCEPTIONS
There is already a sizable body of research on pedagogical implications

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

Joan Garfield and Andrew Ahlgren

51

of students' misconceptions in science, and some of the insight accumulating

from it is relevant to our current inquiry. We cite below several studies that
make particular points about the persistence of preconceptions despite instruction, the lack of correspondence between understanding of a topic and
belief in it, developmental stages of concept understanding, and attempts to
change misconceptions.
Some recent research in physics education is suggestive of the kind of
difficulties students may have and of how to investigate them (Clement,
1982; diSessa, 1982; Fredette & Lochhead, 1980; McCloskey, 1983;
McCloskey, Washburn, & Felch, 1983; McDermott, 1984). These investigations, which involved children, college physics students (including secondary school teachers in training), and adults with no science training,
show that some misconceptions are quite widespread and can persist in spite
of relevant instruction. Moreover, the evidence shows that misconceptions
may exist despite an individual's ability to use appropriate terminology or
to answer questions correctly on a typical test.
Studies of misconceptions of gravity (Sneider & Pulos, 1983) are particularly interesting because they have led to a developmental scheme of stages
through which children appear to progress in an invariant fashion. By recognizing the stage at which a learner is thinking, the researchers suggest,
teachers may be able to tailor instruction to be optimally helpful in making
the transition to the next stage.
There have been some projects directed specifically at changing misconceptions in physics (e.g., Champagne, Gunstone, & Klopfer, 1983; Hewson
& Hewson, 1983; Posner, Hewson, Gertzog, & Strike, 1982). Success has
been elusive. Some important lessons may be drawn from Minstrell (1982),
who attributed his moderate success in changing students' conceptions to
(a) a classroom setting that encourages expression of beliefs without criticism, (b) an emphasis on observation rather than authority, (c) juxtaposition
of several different demonstrations of a phenomenon, and (d) encouragement to identify simple and consistent explanations.
Similar studies have begun on misconceptions in biology, as exemplified
by an excellent study of misconceptions about evolution. Bishop and Anderson (1986) report that number of years of high school and college biology
made no difference in prevalence of misconceptions in college underclassmen-80% of whom had serious misconceptions about fundamentals of
natural selection. This study also testifies to the resilience of misconceptions:
Even after a college biology course in which the misconceptions were systematically challenged, only 50% of the students had relinquished them by
the end of the course. Moreover, students' understanding of natural selection was not correlated with their belief in it-either before or after the
course. It is important to note that Bishop and Anderson also developed a
diagnostic test for identifying students' difficulties, to serve as a basis for
designing instruction.

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

52

LearningStatistics
MISCONCEPTIONS
IN STATISTICAL
REASONING

Identifying Misconceptions
Most of the research on inappropriate statistical reasoning has been done
not by educators, but by psychologists. Piaget and Inhelder (1975) are often
cited for initiating the developmental research that has helped reveal difficulties students have with theoretical conceptions of probability. They have
been criticized, however, for using a solely classical approach to probability
and for basing their experiments on theoretical concepts and proportional
reasoning-which often appeared to be games comparing fractions, with
little connection to ideas of chance (Kapadia, 1985).
More recently, an area of inquiry referred to as judgment under uncertainty has emerged. Key studies in this area were brought together in a book
edited by Kahneman, Slovic, and Tversky (1982). Several articles address
the idea of representativeness, which refers to varieties of the idea that an
occurrence is probable to the extent that it is "typical." In the combinational
form of this misconception, for example, the probability of a man's being
black and being unemployed is judged to be higher than the probability of
his being black. In the sampling form of the representativeness misconception, small samples are generally believed to be more reliable representatives
of a population than is statistically likely. (For example, a preference for
Brand X over Brand Y in four out of five people would typically be believed
to be clearly indicative of a general preference-although the probability of
getting such an extreme sample of 5 just by chance is 3/8.)
Other types of judgment analyzed include availability (a heuristic that
judges a categorization is probable to the extent that instances of it can be
easily brought to mind), and the inference of causality from correlation.
Another relevant topic is covariation and control, in which a central finding
is that people are not likely to spot a correlation unless they expect it-and
if they do expect it, they tend to see one even if it is not there. The clear
story that runs through the articles in the Kahneman et al. (1982) book is
that inappropriate reasoning is (a) widespread and persistent, (b) similar at
all age levels, (c) found even among experienced researchers, and (d) quite
difficult to change.
There has recently been a good deal of research by educators exploring
the probabilistic intuitions of students from elementary through college
level. Much of this research is being done in Europe, in connection with the
Centre for Statistical Education in Sheffield, England, or with the Institut
fiir Didaktik der Mathematik der Universitit Bielefeld, in the Federal Republic of Germany. Results of several studies around the world were presented at the First International Conference on Teaching Statistics, held in
Sheffield, England. Among these are preliminary attempts to describe stages
in the development of probabilistic thinking (Falk, 1983; Fischbein & Gazit,
1983; Green, 1983).

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

Joan Garfieldand AndrewAhlgren

53

Misconceptions Involving Descriptive Statistics

Less extensive attention has been given by psychological and educational
researchers to putatively simple statistical ideas such as distribution, average, sample, and randomness. But there is evidence that conceptual difficulties abound for these topics, too. Indeed, investigators in the Quantitative
Literacy Project have discovered, in beginning with the activity of constructing a simple "line plot," that many junior high school students do not yet
understand what a number line is and that even simple tables of data are
incomprehensible to many students (J. Landwehr, personal communication,
10 January 1985). Pollatsek, Lima, and Well (1981) report on students'
difficulties in understanding the need to weight data in computing a mean.
College students asked to combine two grade point averages that were based
on different numbers of courses into a single average were unable to do the
task correctly. Pollatsek et al. believe that "for many students dealing with
the mean is a computational rather than a conceptual act" (p. 191) and that
knowledge "of a computational rule not only does not imply any real understanding of the basic underlying concept, but may actually inhibit the
acquisition of more adequate (relational) understanding" (p. 202).
In surveying the statistical ideas of mathematically able senior high school
students, Johnson (1985) found that most students regard an average as the
usual or typical value. (For example, when asked about the usefulness of
the average temperature for a city, many responded that it would tell you
what to wear if you went there.)
Attempts to Change Misconceptions
The literature that describes misconceptions about statistics and probability is much more extensive than literature on what can practically be done
to ameliorate them. Fischhoff (1982) reviewed 40 articles in a subfield he
calls debiasing theory, which is devoted to explicit attempts to give professional adults some heuristics that will help to override their faulty intuitions.
He concluded that the debiasing attempts have had very limited success.
Nisbett, Krantz, Jepson, and Kunda (1983) have attempted to explain why
people develop misconceptions and maintain them even with exposure to
contradictory experience. They report that subjects can be taught to give
better answers to probabilistic textbook problems but admit that they are
unsure about whether the subjects are using a formal model to do so and
about the likelihood of transfer to the real world.
There are, however, reports of success. Shaughnessy (1977) reported that
his activity-based course and small-group work helped college students overcome some misconceptions in representativeness and availability. Mevarech
(1983), having found that high school students err in solving problems
about means because they believe that means have the same properties as
simple numbers, reported that it is helpful to provide corrective-feedback

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

54

LearningStatistics

instruction. Ojemann, Maxey, and Snider (1965a, 1965b, 1966) claimed to

have successfully trained children at the fifth-grade level to make significant
improvements in making probability judgments. Fischbein, Pampu, and
Minzat (1970) claimed that, after instruction, fourth-grade students were
able to think accurately about probability. Fischbein and Gazit (1984) made
a similar claim for students in Grades 6 and 7.
Because these studies assessed the accuracy of judgments without probing
the students' means of arriving at them, it is not clear to what extent the
students' ideas of probability changed and to what extent they merely
learned rules. Indeed, Fischbein and Gazit (1984) reported that the children
appeared to do much better on some probability tasks than they were able
to do on proportional reasoning-which one might expect to be requisite
for understanding probability calculations. When students were given a test
of transfer to intuitive ideas about probability, the instructed group of
students did distinctly worse on some parts than the uninstructed group.
Fischbein and Gazit hypothesized that "by emphasizing (via systematic instruction) specific probability viewpoints and procedures one may disturb
the subjects' proportional reasoning, still fragile in many adolescents"
(p. 23).
In a review of selected psychological and pedagogical research studies,
Hawkins and Kapadia (1984) identified key questions concerning how children understand probability. These questions included (a) What conceptions
of probability do children of various ages have? (b) How might these conceptions be changed? (c) Are there optimum teaching and learning techniques? Hawkins and Kapadia believed that because of differences in terminology or research methods used, research has tended to generate debate
rather than answer these questions. In their thorough discussion of classroom implications of the research on children's ideas of probability, they
cited a nationwide study in the United Kingdom in which 11-year-olds first
predicted, and then made a trial of, the number of heads in 12 tosses of a
coin and the pattern of results in 24 rolls of a die. Hawkins and Kapadia
quoted this passage from the study report:
In general the results show a disparity between the quality of predictions and the justifications
given for them. The vast majority of pupils gave mathematically sound predictions. However,
when they found that these predictions were not confirmed by the data, many pupils reverted
to past experience, hunches or cynicism .... Pupils become confused, either clinging to what
they believed must be the mathematically correct answer despite contrary results, or abandoning theory altogether. (p. 363)

Hawkins and Kapadia questioned whether the children's original answers

revealed anything of their thinking about probability.
These reports of attempts at instruction and assessment raise the important issue of the relative merits of retraining students' intuition about probability-say, by extensive exposure to probabilistic simulations on a computer or, on the other hand, by giving them rules for improving their
answers. It is likely that the real issue will be how to optimize the interplay

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

Joan Garfieldand AndrewAhlgren

55

of experience and rules. The design of optimal instruction is complicated

further by the possibility of instruction's having unexpected side effects that
may even interfere with the development of students' understanding-as
Johnson (1985) has found.
How Students Think
A major reason for the limited success of remediation is that the research,
although rich in detail on the incidence of misjudgment, has not attended
to underlying psychological mechanisms. A number of researchers are beginning to produce useful information about how misconceptions operate
that should be useful in designing remediation.
Growing attention to the contextual aspects of probability judgments
(Borgida & DeBono, 1985; Morier & Borgida, 1984; Nisbett et al., 1983)
suggests that an important factor in misjudgment is a misperception of the
question being asked. A tentative conclusion is that considerable advantage
in comprehension may be realized by careful attention to the way in which
questions are posed. For example, when asked what the probability is that
a particular attractive, well-dressed woman is a fashion model (BeythMarom & Dekel, 1983), students typically give an answer of about 70%,
and they give about that same value for the probability that she is an actress
or the probability that she is a cosmetics distributor. Instead of answering
with a probability about the woman's profession, the students apparently
are estimating the probability that a fashion model (or actress or cosmetics
distributor) would be attractive and well dressed. Insisting that the sum of
probabilities must not exceed 1.0 (what Tversky and Kahneman call coherence) may therefore only be confusing to students. It is possible that a great
deal of research, in focusing on the correctness of answers, has missed the
subjects' perception of what the question was-and so misestimated the
subjects' reasoning.
Tversky and Kahneman (1982) have characterized representativeness and
availability not only as classifications of a number of misjudgment phenomena but as general features underlying intuitive reasoning. As appealing as
these constructs are logically, it remains to be shown that they validly
describe thinking. There is evidence both pro and con in some particularly
interesting interview studies of college students' thinking when answering
statistical questions (e.g., Hardiman, Well, & Pollatsek, 1984; Konold,
1983; Pollatsek, Lima, & Well, 1981). The useful integration of research
and instruction is evident in a recent article (Pollatsek, Konold, Well, &
Lima, 1984) on textbook attempts to ameliorate the "gambler's fallacy" by
refuting students' belief in "active balancing," that is, that fluctuation in a
random series will tend to be balanced by an opposite fluctuation:
Our research suggests that such an approach is unfruitful because subjects do not have an
incorrect process mechanism; indeed they have virtually no mechanistic way to think about
random samples. To refute active balancing is to refute a belief that students actually do not

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

56

LearningStatistics

have, and this may confuse them. Since students' actual heuristic, representativeness, is so
different in form from the appropriate mechanistic belief, it may not be easy to effect any
lasting change in students' beliefs about random samples. (p. 400)

Findings like these among college students imply the need for great caution in making objectives for instruction in high school. More important,
however, is the opportunity that the interview methods provide for testing
our propositions of what is going on in students' minds. For example,
Konold (1983) inferred from in-depth interviews that students typically have
an "outcome orientation" model of probability that may be more fundamental than the representativeness model. The outcome orientation involves
making yes-or-no decisions about single events. Konold compared the representativeness and outcome-orientation models in college students' responses to novel questions. For example, when students were asked what
the outcome would be of selecting six marbles from a population that was
5/6 black and 1/6 white, most students thought it more likely that all six
would be black (i.e., a black marble is the most likely outcome on every
single trial) than that one would be white (the most representative result for
the series).
In the same study, Konold found that when students observed frequencies
in repeated trials, they inferred numerical probabilities by using an indirect,
two-stage process: In the first stage, the observed frequencies led to a feeling
of confidence; if a numerical representation was demanded of them, they
then generated a numerical "probability" from the feeling-not from the
frequencies. If true, this finding would imply considerable circumspectness
in using simulations to change the way that students think about probability. Such in-depth exploratory research should be extended to other basic
topics and to students at other levels of education.
IMPLICATIONS
FOR FUTURERESEARCH
The declared importance of introducing statistical concepts into the
school curriculum, together with our limited knowledge about cognitive
development, mathematics learning in general, and misconceptions in probability and statistics, indicates that the time is ripe for an intensive and
coherent assault on the difficulty of learning basic concepts in stochastics.
Indeed, the creation of the International Study Group on Probability and
Statistics Concepts (ASA-NCTM Joint Committee, 1985) and international
conferences on teaching and learning stochastics (Grey, Holmes, Barnett,
& Constable, 1983) indicate that the learning of stochastics is an area of
major interest and activity. Many researchers in education and psychology
have explored parts of the misconceptions picture. Some researchers have
attempted to design specific remedial treatments, and many insightful and
creative teachers have suggested and tried improved techniques of instruction. Accounts of these continue to appear in the journal TeachingStatistics
and the newsletter The Statistics TeacherNetwork.

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

Joan Garfieldand AndrewAhlgren

57

Although strong arguments have been made that students learn best when
instruction is couched in the context of students' "real world" knowledge
(Freudenthal, 1973; Kapadia, 1985; Roth, 1985), there is still only a little
published research on the effectiveness of this approach or any other. This
lack of research is perhaps due, as Bentz and Borovcnik (1985) believe, to
the difficulty of conducting this type of empirical research. Bentz and Borovcnik provide a catalog of problems that have limited the interpretation
of empirical research on probabilistic concepts.
A related problem is a lack of research on the design and use of instruments to measure statistical understanding (Boveda, 1975; Chervany et al.,
1977; Garfield, 1981). A few instruments have been designed to measure
students' attitudes and anxiety toward statistics (Roberts & Saxe, 1982;
Wise, 1985), and some research has appeared that shows the role of factors
influencing general achievement in a statistics course (Harvey, Planke, &
Wise, 1985). Green (1983) has recently developed a test for children ages
11 to 16 that measures three levels of attainment of probability concepts.
The literature makes it clear that far more research has been done on the
psychology of probability than on other statistical concepts. In spite of this
research, however, teaching a conceptual grasp of probability still appears
to be a very difficult task, fraught with ambiguity and illusion. Accordingly,
we make the pragmatic recommendation for two research efforts that would
proceed in parallel: one that continues to explore the means to induce valid
conceptions of probability, and one that explores how useful ideas of statistical inference can be taught independently of technically correct probability. The interpretation of many statistical results must deal in some way
with whether or not we are being fooled by random variation, but it may
be possible to rely on simulation to make semiquantitative arguments. This
view entails accepting students' natural subjectivist view and endeavoring
to get them to draw increasingly on theoretical and frequency information
to set more appropriate belief levels.
As far as goals for instruction go, this view suggests a moratorium on the
typical organization of statistics instruction: (a) descriptive statistics, (b)
probability, and (c) inferential statistics. The intrusion of technical probability issues that are not likely to be understood will stall the learning
process-and leave a distaste that could compromise subsequent instruction
as well. This view is consistent with at least the beginning of the Quantitative
Literacy Project curriculum, in which qualitative approaches are used as
much as possible, computation is minimized, and probabilities are handled
in large part by simulation. The view may not be consistent, on the other
hand, with some of the ambitious goals for including probability in the K12 curriculum, foreshadowed in preliminary drafts of the ASA-NCTM Joint
Committee's (1985) preliminary guidelines. What is needed, however, is not
debate but research.
The difficulty of learning probability and statistics seems not to be a

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

58

Learning Statistics

simple problem that can be ameliorated by simple remedies. The misconceptions already identified are not isolated errors of information or arbitrary
habits of thinking. It seems, rather, that misconceptions are part of a way
of thinking about events that is deeply rooted in most people, either as
learned parts of our culture or (in the extreme) even as brain functions
arising from natural selection in a simpler time. Indeed, one direction for
research would be in social anthropology-to attempt to deduce "misconceptions" from what would have been useful modes of thinking in prehistoric settings. The misconception problem is aggravated by difficulties in
quantitative thinking in general and difficulties in hypothetical reasoning.
These cognitive problems are complicated still further by affective obstacles-faintness of motivation for learning what is believed to be a useless,
forbidding, and even deceptive topic.
A complete research assault on difficulties in learning stochastics would
require attention to all these aspects of the problem. In reviewing the research on instruction in mathematics and science, Resnick (1983) encouraged collaboration between cognitive psychologists and discipline specialists
to improve precollege instruction in mathematics and science. She recommended that improvement begin early, emphasize qualitative reasoning,
build on what students already know, and confront naive intuition directly.
In addition, we believe it will be essential to undertake longitudinal studies
of how individual students actually develop in stochastic sophistication.
Although the accumulation of research findings from diverse current endeavors may eventually yield a substantial understanding of how to teach
probability and statistics better, we would like to add the suggestion that
progress will be realized more surely and efficiently by concentrated and
cooperative projects at single sites, where there can be ongoing interaction
among researchers from the several relevant fields. The sooner such projects
begin the better, because once schools implement the curricula being developed, their commitment to these materials may make more optimal approaches difficult to sell. (Or, if the new curricula do not work adequately,
the schools may develop resistance to stochastics altogether.) Only when
we understand how people learn stochastics concepts will we be able to
prescribe adequately those learning experiences that will be effective.
REFERENCES
American Association for the Advancement of Science (1974). Science-A process approach
II. Nashua, NH: Delta Education.
American Statistical Association-National Council of Teachers of Mathematics Joint Committee on the Curriculum in Statistics and Probability (1985). Teachingof statistics within
the K-12 mathematics curriculum. Washington, DC: American Statistical Association.
Bell, A., Low, B., & Kilpatrick, J. (Eds.). (1985). Theory,researchand practice in mathematical
education. Nottingham, UK: Shell Centre for Mathematical Education.
Behr, M., Lesh, R., Post, T., & Silver, E. (1983). Rational number concepts. In R. Lesh (Ed.),
Acquisition of mathematical concepts and processes (pp. 91-126). New York: Academic
Press.

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

Joan Garfield and Andrew Ahlgren

59

Bentz, H.-J., & Borovcnik, M. G. (1985). Some remarks on empirical investigations on probability and statistics concepts. In A. Bell, B. Low, & J. Kilpatrick (Eds.), Theory, research
and practice in mathematical education (pp. 266-273). Nottingham, UK: Shell Centre for
Mathematical Education.
Beyth-Marom, R., & Dekel, S. (1983). A curriculum to improve thinking under uncertainty.
Instructional Science, 12, 67-82.
Bialaszewski, D. (1981). Statistics, computer packages and the real world. Improving College
and University Teaching,29, 162-165.
Bishop, B. A., & Anderson, C. W. (1986). Student conceptions of natural selection and its role
in evolution. (Research Series No. 165). East Lansing: Michigan State University, Institute
for Research on Teaching.
Borgida, E., & DeBono, K. (1985). Hypothesis-testing and the role of expert knowledge. Manuscript submitted for publication.
Borovcnik, M. (1985). The role of descriptive statistics. In A. Bell, B. Low, & J. Kilpatrick
(Eds.), Theory, researchand practice in mathematical education (pp. 285-292). Nottingham,
UK: Shell Centre for Mathematical Education.
Boruch, R. F., Nowakowski, J., Zawojewski, J., & Goldberg, E. (1984). Coupling cultures:
Teachingstatistics across disciplines in middle and secondary school. Unpublished manuscript,
Northwestern University, Center for Statistics and Probability, Evanston, IL.
Boveda, A. (1975). Diagnosis of deficiencies in the understanding of concepts of statistics.
Dissertation Abstracts International, 37, 188A. (University Microfilms No. 76-14,724)
Cantu, L. L., & Herron, J. D. (1978). Concrete and formal Piagetian stages and science concept
attainment. Journal of Research in Science Teaching,15, 135-143.
Carpenter, T. P., Corbitt, M. K., & Kepner, H. S., Jr. (1981). What are the chances of your
students knowing probability? Mathematics Teacher,74, 342-345.
Carpenter, T. P., Lindquist, M. M., Matthews, W., & Silver, E. (1983). Results of the third
NAEP mathematics assessment: Secondary school. Mathematics Teacher,76, 652-659.
Champagne, A. B., Gunstone, R. F., & Klopfer, L. E. (1983, April). Effecting changes in
cognitive structures among physics students. Paper presented at the meeting of the American
Educational Research Association, Montreal.
Chervany, N. L., Collier, R. D., Fienberg, S., & Johnson, P. (1977). A framework for the
development of measurement instruments for evaluating the introductory statistics course.
American Statistician, 31(1), 17-23.
Clement, J. (1982). Students' preconceptions in introductory mechanics. American Journal of
Physics, 50, 66-71.
Collis, B. (1983). Simulation and the microcomputer: An approach to teaching probability.
Mathematics Teacher,75, 584-587.
Conference Board of the Mathematical Sciences. (1982). New goals for mathematical sciences
education. Washington, DC: Author.
Conference Board of the Mathematical Sciences. (1983). The mathematicalsciences curriculum,
K-12: What is still fundamental and what is not. Washington, DC: Author.
DiSessa, A.A. (1982). Unlearning Aristotelian physics: A study of knowledge-based learning.
Cognitive Science, 6, 37-75.
Doane, D. (1985). Exploring statistics with the IBM PC. Reading, MA: Addison-Wesley.
Duchastel, P. C. (1974). Computer applications and instructional innovation: A case study in
the teaching of statistics. International Journal of Mathematics Education in Science and
Technology,5, 607-616.
Duncan, D. R., & Litwiller, B. H. (1981). Randomness, normality, and hypothesis testing:
Experiences for the statistics class. Mathematics Teacher,74, 368-374.
Edgar, S. M. (1973, June). Teachingstatistics while simultaneously saving time, chalk, etc. Paper
presented at the Conference on Computers in the Undergraduate Curricula, Claremont, CA.
(ERIC Document Reproduction Service No. ED 079 993)
Educational Services. (1968). Elementary science study. St. Louis: Webster/McGraw-Hill.

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

60

Learning Statistics

Elzey, F. F. (1985). Introductory statistics: A microcomputer approach. Belmont, CA: Brooks/

Cole.
Ernest, P. (1984). Introducing the concept of probability. Mathematics Teacher,77, 524-525.
Falk, R. (1983). Choice behavior in probabilistic situations. In D. R. Grey, P. Holmes, V.
Barnett, & G. M. Constable (Eds.), Proceedings of the First International Conference on
TeachingStatistics (pp. 714-727). Sheffield, UK: Teaching Statistics Trust.
Fey, J. T. (Ed.). (1984). Computing and mathematics: The impact on secondary school curricula.
Reston, VA: National Council of Teachers of Mathematics.
Fischbein, E. (1975). The intuitive sources of probabilistic thinking in children. Dordrecht, The
Netherlands: D. Reidel.
Fischbein, E., & Gazit, A. (1983). Does the teaching of probability improve probabilistic
intuitions? In D. R. Grey, P. Holmes, V. Barnett, & G. M. Constable (Eds.), Proceedings of
the First International Conferenceon TeachingStatistics (pp. 738-753). Sheffield, UK: Teaching Statistics Trust.
Fischbein, E., & Gazit, A. (1984). Does the teaching of probability improve probabilistic
intuitions? Educational Studies in Mathematics, 15, 1-24.
Fischbein, E., Pampu, I., & Minzat, I. (1970). Comparison of ratios and the chance concept
in children. Child Development, 41, 377-389.
Fischhoff, B. (1982). Debiasing. In D. Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment
under uncertainty: Heuristics and biases (pp. 422-444). Cambridge, UK: Cambridge University Press.
Fredette, N., & Lochhead, J. (1980). Student conceptions of simple circuits. Physics Teacher,
18, 194-198.
Freudenthal, H. (1973). Mathematics as an educational task. Dordrecht, The Netherlands: D.
Reidel.
Garfield, J. (1981). An investigation of factors influencing student attainment of statistical
competence. (Doctoral dissertation, University of Minnesota, 1981). Dissertation Abstracts
International, 42, 5064A.
Gnanadesikan, M., Scheaffer, R. L., & Swift, J. (1986). The art and techniques of simulation.
Palo Alto, CA: Dale Seymour.
Green, D. R. (1983). A survey of probability concepts in 3000 pupils aged 11-16 years. In D.
R. Grey, P. Holmes, V. Barnett, & G. M. Constable (Eds.), Proceedings of the First International Conference on TeachingStatistics (pp. 766-783). Sheffield, UK: Teaching Statistics
Trust.
Grey, D. R., Holmes, P., Barnett, V., & Constable, G. M. (Eds.). (1983). Proceedings of the
First International Conference on TeachingStatistics. Sheffield, UK: Teaching Statistics Trust.
Hansen, R. S., McCann, J., & Myers, J. L. (1985). Rote versus conceptual emphases in teaching
elementary probability. Journal for Research in Mathematics Education, 16, 364-374.
Hardiman, P., Well, A. D., & Pollatsek, A. (1984). Usefulness of a balance model in understanding the mean. Journal of Educational Psychology, 76, 792-801.
Harvey, A. L., Planke, B. S., & Wise, S.L. (1985, April). The validity of six beliefs about factors
related to statistics achievement. Paper presented at the meeting of the American Educational
Research Association, Chicago.
Hawkins, A. S., & Kapadia, R. (1984). Children's conceptions of probability-A psychological
and pedagogical review. Educational Studies in Mathematics, 15, 349-378.
Herron, J. D. (1978). Piaget in the classroom. Journal of Chemical Education, 55, 165-169.
Hewson, M., & Hewson, P. (1983). Effect of instruction using students' prior knowledge and
conceptual change strategies on science learning. Journalof Researchin Science Teaching,20,
731-843.
Hope, J. A., & Kelly, I. W. (1983). Common difficulties with probabilistic reasoning. Mathematics Teacher,76, 565-570.
Johnson, N. C. (1985). Role of dynamic microcomputer graphics in teaching the meaning of
random sampling and the central limit theorem. (Doctoral dissertation, University of Minnesota, 1985). Dissertation Abstracts International, 47, 455A.

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

Joan Garfield and Andrew Ahlgren

61

Jolliffe, F. R. (1976). A continuous assessment scheme for statistics courses for social scientists.
InternationalJournal of Mathematics Education in Science and Technology,7, 97-103.
Kahneman, D., Slovic, P., & Tversky, A. (Eds.). (1982). Judgment under uncertainty: Heuristics
and biases. Cambridge, UK: Cambridge University Press.
Kalton, G. (1973). Problems and possibilities with teaching introductory statistics to social
scientists. International Journal of Mathematics Education in Science and Technology,4, 716.
Kapadia, R. (1985). A brief survey of research on probabilistic notions. In A. Bell, B. Low, &
J. Kilpatrick (Eds.), Theory, researchand practice in mathematical education (pp. 261-265).
Nottingham, UK: Shell Centre for Mathematical Education.
Kempthorne, 0. (1980). The teaching of statistics: Content versus form. American Statistician.
34(1), 17-21.
Konold, C. E. (1983). Conceptions of probability: Reality between a rock and a hard place.
(Doctoral dissertation, University of Massachusetts, 1983). Dissertation Abstracts International, 43, 4179B.
Landwehr, J., & Watkins, A. E. (1986). Exploring data. Palo Alto, CA: Dale Seymour.
Landwehr, J., Watkins, A. E., & Swift, J. (1987). Exploring surveys: Information from samples.
Palo Alto, CA: Dale Seymour.
McCloskey, M. (1983, April). Intuitive physics. Scientific American, pp. 122-130.
McCloskey, M., Washburn, A., & Felch, L. (1983). Intuitive physics: The straight-down belief
and its origin. Journal of Experimental Psychology, 9, 636-649.
McDermott, L. (1984). Research on conceptual understanding in mechanics. Physics Today,
37, 24-32.
Mevarech, Z. R. (1983). A deep structure model of students' statistical misconceptions. Educational Studies in Mathematics, 14, 415-429.
Milakofsky, L., & Patterson, H. D. (1979). Chemical education and Piaget. Journalof Chemical
Education, 56(2), 87-90.
Minstrell, J. (1982). Conceptual development research in the natural setting of a secondary
school classroom. In M. B. Rowe (Ed.), Science for the 80's (pp. 129-143). Washington,
DC: National Education Association.
Morier, D., & Borgida, E. (1984). The conjunction fallacy: A task specific phenomenon?
Personality and Social Psychology Bulletin, 10, 243-252.
Mosteller, F. (Ed.). (1973). Statistics by example. Reading, MA: Addison-Wesley.
Newman, C. M., Obremski, T. E., & Scheaffer, R. L. (1986). Exploring probability. Palo Alto,
CA: Dale Seymour.
Nisbett, R., Krantz, D., Jepson, C., & Kunda, Z. (1983). The use of statistical thinking in
everyday inductive reasoning. Psychological Review, 90, 339-363.
Noddings, N., Gilbert-MacMillan, K., & Lutz, S. (1980, April). What does an individual gain
in small group mathematical problem solving? Paper presented at the meeting of the American
Educational Research Association, Montreal.
Norusis, M. J. (1986). The SPSS guide to data analysis. Chicago: SPSS.
Nygard, K.E. (1983). A system for generating instructional computer graphics. AEDS Journal,
16, 177-187.
Ojemann, R. H., Maxey, E. J., & Snider, B. C. (1965a). Effects of guided learning experiences
in developing probability concepts at the fifth grade level. Perceptual & Motor Skills, 21,
415-427.
Ojemann, R. H., Maxey, E. J., & Snider, B. C. (1965b). The effect of a program of guided
learning experiences in developing probability concepts at the third grade level. Journal of
Experimental Education, 33, 321-330.
Ojemann, R. H., Maxey, E. J., & Snider, B. C. (1966). Further study of guided learning
experience in developing probability concepts in grade five. Perceptual & Motor Skills, 23,
323-336.
Pereira-Mendoza, L., & Swift, J. (1981). Why teach statistics and probability-A rationale. In

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

62

Learning Statistics

A. P. Shulte (Ed.), Teachingstatistics and probability (1981 Yearbook of the National Council
of Teachers of Mathematics, pp. 1-7). Reston, VA: NCTM.
Piaget, J., & Inhelder, B. (1975). The origin of the idea of chance in children. London: Routledge
& Kegan Paul.
Pollatsek, A., Konold, C. E., Well, A. D., & Lima, S. D. (1984). Beliefs underlying random
sampling. Memory and Cognition, 12, 394-401.
Pollatsek, A., Lima, S., & Well, A. D. (1981). Concept or computation: Students' understanding of the mean. Educational Studies in Mathematics, 12, 191-204.
Posner, G. J., Hewson, B. W., Gertzog, W. A., & Strike, K. A. (1982). Accommodation of a
scientific conception: Toward a theory of conceptual change. Science Education, 66, 211227.
Resnick, L. R. (1983). Mathematics and science learning: A new conception. Science, 220,
477-478.
Richbart, L. A. (1981). Probability and statistics for grades 9-11. Mathematics Teacher,74,
346-348.
Roberts, D. M., & Saxe, J. E. (1982). Validity of a statistics attitude survey: A follow-up study.
Educational and Psychological Measurement, 42, 907-912.
Roth, K. J. (1985, April). Conceptual change learning and student processing of science texts.
Paper presented at the meeting of the American Educational Research Association, Chicago.
Ryan, T., Joiner, B., & Ryan, B. (1976). MINITAB student handbook. North Scituate, MA:
Duxbury Press.
Scheaffer, R. L. (1984). The role of statistics in revitalizing precollege mathematics and science
education. In Proceedings of the Section on Statistical Education (pp. 19-21). Washington,
DC: American Statistical Association.
Schoenfeld, A. H. (1985, March). Understanding and teaching the nature of mathematical
thinking. Paper presented at the University of Chicago School Mathematics Project International Conference on Mathematics Education, Chicago.
Shaughnessy, J. M. (1977). Misconceptions of probability: An experiment with a small-group,
activity-based, model building approach to introductory probability at the college level.
Educational Studies in Mathematics, 8, 295-316.
Shaughnessy, J. M. (1981). Misconceptions of probability: From systematic errors to systematic experiments and decisions. In A. P. Shulte (Ed.), Teachingstatistics and probability (1981
Yearbook of the National Council of Teachers of Mathematics, pp. 90-99). Reston, VA:
NCTM.
Shufelt, G. (Ed.). (1983). The agenda in action (1983 Yearbook of the National Council of
Teachers of Mathematics). Reston, VA: NCTM.
Shulte, A. P. (Ed.). (1981). Teachingstatistics and probability (1981 Yearbook of the National
Council of Teachers of Mathematics). Reston, VA: NCTM.
Smith, P. J. (1978). Piaget in high school instruction. Journal of Chemical Education, 55, 115118.
Sneider, C., & Pulos, S. (1983). Children's cosmographics: Understanding the earth's shape
and gravity. Science Education, 67, 205-221.
Stockburger, D. W. (1982). Evaluation of three simulation exercises in an introductory statistics
course. Contemporary Educational Psychology, 7, 365-370.
Stonewater, J. K., & Stonewater, B. B. (1984). Teaching problem solving: Implications from
cognitive development research. AAHE Bulletin, 36(6), 7-10.
Stroup, D. F. (1984). The statistician and the pedagogical monster: Characteristicsof effective
instructors of large statistics classes. In Proceedings of the Section on Statistical Education
(pp. 1-7). Washington, DC: American Statistical Association.
Swift, J. (1982, December). The first international conference on teaching statistics. Statistics
TeacherNetwork, pp. 1-2.
Swift, J. (1983). Challenges for enriching the curriculum: Statistics and probability. Mathematics Teacher,76, 268-269.

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions

Joan Garfield and Andrew Ahlgren

63

Thomas, D. A. (1984). Understanding the central limit theorem. Mathematics Teacher,77, 542543.
Tversky, A., & Kahneman, D. (1982). Causal schemas in judgments under uncertainty. In D.
Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment under uncertainty: Heuristics and biases
(pp. 117-128). Cambridge, UK: Cambridge University Press.
University of Chicago School Mathematics Project. (1985, Autumn). Chicago: University of
Chicago, UCSMP.
Urquhart, N. S. (1971). Nonverbal instructional approaches for introductory statistics. American Statistician, 25(2), 20-25.
Vandermeulen, H., & DeWreede, R. E. (1983). The amazing histogram. inCider, 1(11), 5356.
Wise, S. L. (1985, April). The development of a scale measuring attitudes toward statistics.
Paper presented at the meeting of the American Educational Research Association, Chicago.

AUTHORS
JOAN GARFIELD, Assistant Professor, General College, 106 Nicholson Hall, University of
Minnesota, 216 Pillsbury Drive S.E., Minneapolis, MN 55455
ANDREW AHLGREN, Professor (on leave from College of Education, University of Minnesota), American Association for the Advancement of Science, 1333 H Street N.W., Washington, DC 20005

This content downloaded from 163.178.101.228 on Mon, 9 Dec 2013 15:11:06 PM

All use subject to JSTOR Terms and Conditions